Embed Size (px)
Citation preview
EFFECTS OF EVERGLADES RESTORATION ON SUGARCANE FARMING IN
THE EVERGLADES AGRICULTURAL AREA
By
JENNIE MARIA VARELA
A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE
UNIVERSITY OF FLORIDA
2005
Copyright 2005
by
Jennie Maria Varela
This thesis is dedicated to my parents, Carlos and Janet, and my sister, Carmen.
iv
ACKNOWLEDGMENTS
I extend my deepest gratitude to my supervisory committee chair, Dr. Donna Lee,
and committee members, Dr. Clyde Kiker and Dr. Alan Hodges, for their guidance and
assistance over the course of my thesis research. I am very thankful to Dr. Rick Weldon
for his advisement regarding the analysis presented in this document and to Barry Glaz
and Forest Izuno for their personal cooperation and contributions to this project.
I also wish to express my appreciation to the faculty and staff members in the Food
and Resource Economics Department, and to my fellow graduate students for their
support and encouragement throughout my course of study.
Finally, I would like to thank my extended family and friends for their constant
support and unwavering confidence. I am especially grateful to the community of St.
Augustine Church and Catholic Student Center whose friendship, love, and prayers made
possible my success at the University of Florida.
v
TABLE OF CONTENTS page
ACKNOWLEDGMENTS ................................................................................................. iv
LIST OF TABLES............................................................................................................ vii
LIST OF FIGURES ......................................................................................................... viii
ABSTRACT....................................................................................................................... ix
CHAPTER
1 INTRODUCTION ........................................................................................................1
Geography and Land Use .............................................................................................2 Economic Characteristics: EAA...................................................................................5 Restoration and Conservation History..........................................................................5 Comprehensive Everglades Restoration Plan...............................................................7 Focus of Present Work..................................................................................................8
Problem Statement.................................................................................................8 Hypotheses ............................................................................................................8
Maintaining a Higher Water Table Lowers Average Sugarcane Production for an EAA Farm......................................................................8
The EAA Sugarcane Operation Will Experience a Reduction in Profit Under the Changed Water Conditions. .......................................................9
Research Objectives ..............................................................................................9
2 PRODUCTION THEORY AND ITS APPLICATION TO FLORIDA AGRICULTURE ........................................................................................................11
Theory of the Firm......................................................................................................11 Interrelationships of Economic and Agronomic Concepts.........................................12 Diminishing Returns...................................................................................................15 Modeling Production ..................................................................................................16 Modeling Production and Cost ...................................................................................20 Economics of Water Use ............................................................................................20 South Florida Agriculture and Ecosystem Restoration ..............................................21 Sugarcane response to high water tables and flooding...............................................22
3 METHODOLOGY .....................................................................................................26
vi
Introduction and Overview of Analysis......................................................................26 Agronomic Model.......................................................................................................27 Rainfall Model ............................................................................................................29
4 DATA SOURCES ......................................................................................................31
Empirical Research on Sugarcane Response..............................................................31 Water Table and Flooding Conditions........................................................................32 Historical Production ..................................................................................................34 Climatic Data ..............................................................................................................35 Costs of Production.....................................................................................................36 Sugarcane Prices.........................................................................................................37
5 RESULTS AND DISCUSSION.................................................................................39
Empirical Model Results ............................................................................................39 Rainfall Model Results ...............................................................................................41 Comparison of Model Scenarios ................................................................................42 Evaluation of Hypotheses ...........................................................................................43
Maintaining a Higher Water Table Lowers Average Sugarcane Production for an EAA Farm. ..................................................................................................43
The EAA Sugarcane Operation Will Experience a Reduction in Profit Under the Changed Water Conditions. .......................................................................44
6 SUMMARY AND CONCLUSIONS.........................................................................45
Summary.....................................................................................................................45 Conclusions.................................................................................................................45 Implications for Future Analysis ................................................................................48
APPENDIX
A SIMULATION OUTPUT FOR SCENARIO 1..........................................................50
B SIMULATION OUTPUT FOR SCENARIO 2..........................................................54
C SIMULATION OUTPUT FOR SCENARIO 3..........................................................57
D SIMULATION OUTPUT FOR SCENARIO 4..........................................................60
LIST OF REFERENCES...................................................................................................63
BIOGRAPHICAL SKETCH .............................................................................................66
vii
LIST OF TABLES
Table page 3-1. Key output variables and probability distributions for empirical model....................28
4-1. Total monthly rainfall in inches for the EAA 1979-2000 . .......................................35
4-2. Average EAA rainfall. ................................................................................................36
4-3. Florida sugarcane production expenses......................................................................37
5-1. Summary statistics for simulated model output, profit in dollars..............................40
5-2. Summary statistics for simulated model output, yield in tons per acre.....................41
viii
LIST OF FIGURES
Figure page 1-1. Original Everglades, surrounding wetlands, and south Florida watershed. ................3
1-2. Current map of the Everglades region, including the Everglades Agricultural Area. ...........................................................................................................................4
4-1. Distribution of water level for Hendry County, FL 1977-1995. ................................33
4-2: Total sugarcane production for Hendry County, FL from 1994-2004. ......................34
4-3: Season average price: sugarcane for sugar and seed 1980-2003................................38
ix
Abstract of Thesis Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Science
EFFECTS OF EVERGLADES RESTORATION ON SUGARCANE FARMING IN THE EVERGLADES AGRICULTURAL AREA
By
Jennie Maria Varela
December 2005
Chair: Donna Lee Major Department: Food and Resource Economics
Sugarcane production is a $700 million business in the Everglades Agricultural
Area. Beginning in the 1920s, this portion of the Everglades region of South Florida was
drained, leaving rich muck soils to be used by agriculture. Productivity led to
diminishing water quality until, in 1994, the Everglades Forever Act called for the
wetland to be restored. As part of the larger Comprehensive Everglades Restoration
Project (CERP), the current drainage system in Florida Everglades is being altered or
removed and best management practices require that less water be drained out of the area
to reduce phosphorus loads. It is expected that maintaining a higher water table lowers
sugarcane production for an EAA farm and that the EAA sugarcane operation will
experience economic losses under the changed water conditions.
The analysis assumed a hypothetical 640-acre sugarcane farm with a high level of
management, operating to maximize profit, and independent of processors. Using an
approach based on agronomic yield models, this analysis estimated yield and profit
x
change for this “typical” farm under four scenarios. Functions from a 2002 study on
sugarcane cultivar response to water tables were used to determine an expected yield
function that included a parameter for flood events along with historical distributions for
water table depth. That function, along with acreage, overhead cost, variable cost, and
price, was then simulated using Simetar, varying the input factors.
The mean values for yield changes, and consequently for profit changes, in the
simulations were significantly different across the baseline and post-restoration scenarios.
A zero-profit estimation found that water table depth of 31.83cm with 6 flood cycles
resulted in a yield of 31.1 tons per acre, just a 7 % difference from the historical mean.
All post-restoration simulations resulted in average losses over the estimated period
including a decrease in yield to approximately 27 and 24 tons per acre. If basing
decisions on the mean values of these scenarios, one would expect that the 640-acre farm,
even keeping all acreage in production, would see an average loss of up to $135,000 for
the year.
1
CHAPTER 1 INTRODUCTION
Agriculture is one of Florida’s largest industries and a major sector of the economy.
This industry alone accounts for over six billion dollars annually. Perhaps most well
known commodities are citrus and sugarcane, vegetables, berries and melons While
farmland may not be as expansive as in other states, much of it is in use for these types of
high value crop production. Much of this production takes place in south Florida, which
includes the Everglades Agricultural Area (EAA).
For many, the Florida environment is just as valuable as its industries. Florida has
diverse ecosystems including lakes and rivers. Citizens and lawmakers alike have
worked to restore fragile wetlands and greenways. Programs such as “Florida Forever”
set aside vulnerable parcels of land so that natural areas can be preserved. Public
awareness has influenced initiatives for protected wildlife, sensitive land, and water
resources in many forms, but perhaps none greater than the task of restoring the Florida
Everglades. With strong citizen support, state and federal agencies came together to
develop the Comprehensive Everglades Restoration Plan (CERP). It is this multi-stage
effort that is bringing about great challenges in balancing the interests of producers,
environmentalists, and developers.
It was the establishment of agriculture that motivated the creation of a system for
flood control and began the series of changes in the Everglades area. Later on,
population booms and urban growth further changed the landscape of the state. This
expansion put further pressures on water quality and management. Construction and land
2
development show no signs of slowing, and thus water management will continue to be
an issue for the foreseeable future.
Due to these changes, current producers have a number of immediate concerns:
keeping production profitable, adjusting their practices to meet environmental standards,
and making decisions with an uncertain future in their industry. The EAA is just a small
piece of the larger picture. This area represents jut one sector of one of the most complex
and long-range wetland restoration projects ever undertaken. Within this area, the
primary concerns are maintaining flood control, while also ensuring water availability,
and controlling runoff into the Everglades Protection Area.
The case is unique as the EAA falls within a particular watershed, is home to crops
that may not be produced in many areas, and has been subject to specific water
management measures for so many years. However, Florida is not the only state trying to
balance agricultural and environmental interests, nor is South Florida the only region
struggling to manage development and water needs as well as wanting to preserve as
much of the natural beauty as possible. As these challenges are approached, the results of
these programs will surely serve as indicators for future projects around the country.
Geography and Land Use
The Florida Everglades region is historically known as one of the most unique and
productive ecosystems in the world . Marjory Stoneman Douglas describes the region in
her 1947 book, Everglades: River of Grass:
The grass and the water together make the river as simple as it is unique. There is no other river like it. Yet within that simplicity, enclosed within the river and bordering and intruding on it from each side, there is a subtlety and diversity, a crowd of changing forms, of thrusting teeming life. All that becomes the region of the Everglades.
3
It is the defining ecosystem of South Florida, a hydrological network of saw grass
plains and swamps that once covered nearly three million acres (USGS 2002).
Figure 1-1. Original Everglades, surrounding wetlands, and south Florida watershed. Source: USGS.
From the extensive Everglades marsh, approximately 700,000 acres were drained
to provide rich farmland (Bottcher and Izuno 1994). This area is now known as the
Everglades Agricultural Area (EAA) and sits in Hendry and Palm Beach counties
between Lake Okeechobee to the north and the Everglades Protection Area to the south.
4
Figure 1-2. Current map of the Everglades region, including the Everglades Agricultural Area. Source: IFAS
Agricultural development of the area was made possible by the establishment of a
drainage and irrigation system to regulate the amount of water available, especially
during the wet and dry seasons. The area receives enough annual rainfall to sustain its
crops, however, most of that rainfall comes in June through September, while winter and
spring are dry (EREC 2005). The irregularity of these rainfall patterns makes water
management the EAA=s greatest challenge.
5
The EAA has been able to regulate its water levels by using a system of canals,
pumps, and levees first put in place in the 1920's. Currently, approximately 80% of the
area pumps its excess water into storm water treatment areas and water conservation
areas and a few drainage districts still drain runoff into Lake Okeechobee (FFWP).
While agriculture thrived over the decades, the water quality of the Everglades
diminished with nutrient levels steadily increasing.
Economic Characteristics: EAA
The EAA is still one of the most productive agricultural areas in the country. The
agricultural industries occupying the EAA are responsible for an estimated $1.5 billion of
sales each year (Aillery et al 2001). As of 1997, Palm Beach and Hendry counties had
over 730,000 acres of cropland making up approximately 1200 farms. The leading crops
in production are sugarcane, rice, sod, and winter vegetables. By 2002, the EAA had
about 500,000 acres in production, with 90% of its acreage in Palm Beach County. Over
5,000 people in this area are employed either directly in agriculture or in associated
businesses as reported in the 2000 census (USACE and SFWMD 2003.)
Sugarcane dominates production in the EAA covering 86% of its acreage and
bringing in sales of over $762 million in 2001. Nearly a quarter of domestic sugar is
produced in the region. The EAA also boasts a rice industry with sales of over $9
million. Winter vegetable production is also a profitable practice in the EAA, second only
to sugarcane. Row crops cover over 16,000 acres and represent about 16% of EAA sales
in 2001 (USACE and SFWMD 2003).
Restoration and Conservation History
Regulation of the Everglades region began in 1934 when Congress authorized the
acquisition of land for a park that would preserve natural conditions in south Florida
6
(SFWMD 2004). After thirteen years, President Truman officially dedicated the
Everglades National Park in December of 1947 (ENP), the same year Douglas published
her account of the wetlands.
Florida continued to grow, however, and in order to establish a system for flood
control, agricultural and urban water supply, and preservation of wildlife, Congress in
1948, set forth the Central & Southern Florida (C&SF) Project. Reaching from central
Florida to the Florida Keys, it took the form of canals, levees, storage areas and other
water control structures. The drainage projects for controlling flooding were begun by
the State of Florida and eventually continued by the Corps of Engineers (USACE and
SFWMD 1999). Successful drainage of the area made the land south of Lake
Okeechobee suitable for agricultural development, creating the EAA.
However, this control system altered the natural ecosystem to such an extent, that
even the protected area was being damaged by changes in water flows and the
phosphorus running off of the agricultural lands. In 1991, the State of Florida established
the Douglas Everglades Protection Act (F.S. 373.4592, 1991) which called for a Surface
Water Improvement and Management (SWIM) plan and a change in regulatory
procedures for the EAA. Many felt there was not enough available information to make
such decisions and lawsuits began, delaying the restoration efforts.
In 1994, the Florida state legislature passed the more comprehensive “Everglades
Forever Act” which established new storm water treatment areas and Best Management
Practices (BMP) for the EAA in an effort to restore the natural flow of the Everglades as
well as improving water quality (Bottcher and Izuno 1994). Combining short and long
term projects, this plan includes more research and monitoring in the decision-making
7
process than previous laws had allowed. The state followed up with the Water Resources
Development Act (WRDA) of 1996, authorizing the Corps of Engineers to develop a plan
by 1999 that would encompass all aspects restoration. The Water Resources
Development Act of 2000 set forth the parameters for the final restoration project and
laid out the key points of the comprehensive plan.
Comprehensive Everglades Restoration Plan
The Comprehensive Everglades Restoration Plan (CERP) was the result of the
WRDA of 2000. With a budget of $7.8 billion, the plan includes thirty years of
construction and an additional twenty years of maintenance. There are more than sixty
components of the plan, being carried out by both the SFWMD and the Corps of
Engineers. The costs are to be evenly divided between the state and federal governments.
The primary goals are to develop ecological values as well as increasing economic
and social values. By restoring a more natural hydrological flow, it is expected that water
quality would be improved throughout the area and threatened and endangered species
would also see improvement in their habitat. Under CERP, new water quality treatment
facilities will be established along with over 200,000 acres of reservoirs and wetland
water treatment areas being constructed.
Agricultural concerns have been considered a major part of the plan, because of the
proximity of natural and agricultural areas. In November of 1999, the South Florida
Action Plan for the Applied Behavioral Sciences, drafted to bring socioeconomic
concerns into the restoration planning process brought these concerns to light. A number
of CERP projects take place within or adjacent to current agricultural areas. Though
there are pockets of agriculture spread over south Florida, the area concentrated around
8
Lake Okeechobee, including the EAA, would be the most affected by CERP
implementation (USACE and SFWMD 2001).
Focus of Present Work
This study seeks to provide insight into the role that the Everglades restoration
program will have on the immediate area, specifically, sugarcane farming in the EAA.
EAA farms are crucial to Florida’s agricultural industry and indeed are major
contributors to the greater economy.
Problem Statement
Under CERP, some major drainage canals will be removed and more water is being
retained as part of BMP. In light of these conditions, a sugarcane producer in the EAA is
likely to face higher water tables and longer flood durations, as pumping is limited both
voluntarily and by regulation. To examine these conditions requires looking at many
parts of the situation: options for modeling agricultural production, the relationship
between agriculture and the Everglades restoration effort itself, the economics of water
use, and previous work on sugarcane yields.
By examining the current status of sugarcane production in the EAA and the
current strategies in water management, this study aims to provide an informed picture of
the impacts CERP will have on the agricultural industry in the long run. Looking at the
impacts of water on those crops will assist in filling in some of the gaps in information
Hypotheses
Maintaining a Higher Water Table Lowers Average Sugarcane Production for an EAA Farm.
Though each year brings a season of heavy rain to South Florida, farmers have
been able to manage their risk by planting alternate crops or letting the land lie fallow
9
during the rainy season. Under the restoration plan, it is expected that there will be a
significant difference, as more water is being kept on the land to reduce nutrient loading
in addition to removal of some of the drainage system. Assuming that producers are
currently operating at an optimal level, increasing the available water would hinder yield.
Some crops, such as rice, require flooding and as such may not show a decreased yield.
In the case of sugarcane, lower yield among the current cultivars is anticipated and in
response new cultivars are being bred for water tolerance.
The EAA Sugarcane Operation Will Experience a Reduction in Profit Under the Changed Water Conditions.
Maintaining more water on farmland is a strategy being employed by sugarcane
growers in order to reduce soil subsidence as well as limiting phosphorus loads, major
objectives of the restoration effort. Literature suggests that production is constrained not
only by the amount of water applied, but also by limitations on the amount of water that
can be pumped out of the system under a certain time. The producer, having to choose
which fields to drain, may find that less acreage can be harvested. The consequential
decline in production cuts into the profits of that operation.
Research Objectives
The first objective of the analysis is to provide a descriptive framework for
analyzing changes in sugarcane production for a representative farm in the Everglades
Agricultural Area. This includes determining the current production practices in the area
as well as estimating future production possibilities. In particular, the focus of the work
is to describe the crop response to water management such that a producer would have
knowledge of how changes in those practices could affect his operation. This requires
10
not only gathering information on the crop yield itself, but also determining how much
water is coming in and out of the system.
The other major objective is to gauge the economic impact of the water flow
change. In order to meet this objective, there must first be a determination of future
production. By simulating the production relationship, a range of possibilities can be
outlined. Along with future estimations, there must also be a determination of cost
structure. Once these factors are combined, the resulting analysis can be useful for a
producer in weighing future decisions.
11
CHAPTER 2 PRODUCTION THEORY AND ITS APPLICATION TO FLORIDA AGRICULTURE
Theory of the Firm
As modern economic theory developed, there developed a need for specific
definitions and assumptions. The focus of analysis shifted in the early twentieth century
from industry perspective to that of the individual unit. Coase, specifically, set forth his
Nature of the Firm (1937) in order that continued analysis would have a well-defined and
consistent basis when performed at firm level. At the same time, it was hoped that the
resulting definition would be both realistic and useful within the analytic methods of the
time. Any definition, he stated, would need to conform to the “most powerful economic
instruments” of economics: the ideas of marginal analysis and substitution.
An initial step in creating the basic assumptions is differentiating the firm from
the larger economic system. The basics of the system are well defined and lead the
economist to assume that resources will be allocated based on prices and that the system
will continue working on its own through market transactions. When the perspective
changes to the firm level, however, there is not an internal market to allocate resources.
Instead the firm has to have a coordinator, someone who will be the decision-maker and
direct production. Through this coordinator, the firm avoids the costs that come with
operating a market. The firm is then defined as a system of interactions in which
resources are allocated by a particular “entrepreneur.”
Among the most important characteristics of the theoretical firm are that it has an
upward sloping cost curve and that it will not pay to produce output beyond the point at
12
which marginal cost is equal to marginal revenue. Additionally, the firm will continue to
grow until the marginal cost of creating a transaction within the firm is equal to the cost
of having that transaction in the marketplace.
Alchian and Demsetz (1972) do not refute Coase’s definitions, but take a more
specific look at what constitutes a “classical firm”. They examine the structure and lay
out six major qualifications of a firm. First, it must produce using joint inputs. Second,
these inputs come from a number of input owners. Next, all the contracts for joint inputs
have a party in common. That party must also have the authority to renegotiate the input
contracts independent of each other. Additionally, that party holds claim to the residuals
and finally, must have the right to sell the central contractual residual status.
Similar to Coase, Alchian and Demsetz see the existence of a central decision-
maker as imperative to the existence of the firm. This decision-maker is owner and
employer. They feel that this structure is the result of necessity. That is, that in
attempting to align productivity to the marginal costs of inputs, the most efficient method
is to operate as a classical firm.
Interrelationships of Economic and Agronomic Concepts
Agricultural decision-making involves combining both economic and biological
factors in order to maximize outcome. This combination becomes crucial in examining
yield responses, optimal output, and the overall input-output relationship (Redman and
Allen, 1954). What is considered “optimal” they find, is greatly influenced by the
particular concepts being employed. That is, when constants are changed, other factors
(e.g., the factor-product price ratio) may become more or less important in determining
the most profitable choice.
13
Crop yield is certainly one area in which these two schools of thought intersect.
Agronomic production functions give the economist a quantitative look at the changes in
production under various conditions. Yield is a result of numerous factors, from the plant
itself to the soil in use, to the surrounding climate, to nutritional inputs. Early theorists
concentrated on the “food” of plants: water, nitrogen, earth, fire. As agronomy has
developed however, there is a more clear understanding of plant processes such as
photosynthesis, which incorporates energy into the relationship. Some of these basic
factors, however (water, nitrogen, and other fertilizers) are still the subject of economic
analysis, especially under changing conditions.
The “Law of the Minimum” argued by von Liebig was one of the earliest models
of production and still carries some influence, even if it no longer stands alone. Von
Liebig began with the concept of a minimum factor, the factor of yield that is most
scarce. In this theory, yield will change only when this minimum factor is changed.
Consequently, the production curve would increase at a constant rate until it reached the
limit determined by the minimum factor. At that point, von Liebig’s curve becomes
horizontal, as adding more of the other factors does not change yield.
While von Liebig’s concept resonated with early farm economists, it was not an
especially accurate model of plant response. Later experiments provided data that would
be used to modify the concept, moving away from the idea of a linear relationship with
constant returns to scale.
In trying to improve the production model, Mitscherlich assumed that there was a
maximum yield under ideal conditions and that it was the shortage of any one factor that
14
would cause yield to decrease proportionally. His result was a curve concave to the given
factor axis such that
dy/dx = (A - y) C
where y is the yield while x is the factor in question and A is the maximum yield.
Though an improvement over the von Liebig equation, Mitscherlich’s model still did not
adequately represent possible yields since each factor had its own constant, C and did not
account for factors influencing each other. Baule expanded this model to include variable
growth factors, making the case that yield is dependent upon the interaction of many
factors. Overall yield is then expressed as
Y = (1-10c1x1)(1-10c2x2)…(1-10cnxn)
with c representing the effect of the corresponding x factors.
At the same time, Spillman was developing another estimation of yield. Basing
his model on the expectation that increments of a growth factor could be the terms of a
geometric series. Using the example of fertilizer application, Spillman expressed the
yield relationship as:
Y= A(1-Rx)
In this expression, R would indicate the ratio of increments in yield and suggests a
sigmoid curve. Both Mitscherlich and Stillman agreed that the law if diminishing
increment would not apply once the input quantity was large enough to damage the plant.
Redman and Allen in their overview of these principles raise the concern that
economists must be careful in using data from farm crops on the basis that the functions
drawn from these data are not necessarily true of all plant growth. Fixed factors and
decision-making may be different for separate sets of data and as such, the economist
15
attempts to find an expression of “best fit.” Once such an expression is created, it can
then be used under various scenarios to forecast the profitability of choices.
Diminishing Returns
One of the most relevant parts of the economics/agronomic relationship is the
theory of diminishing returns. This idea was first set forth in the 18th century by Turgot.
In describing expenditures and production, he noted that while returns initially increased,
effects would eventually diminish. Early in the 19th century, Malthus, Ricardo, West, and
Torrens all described the same phenomenon in their publications on land rent. Ricardo
was perhaps most accurate in describing the phenomenon within intensive farming. His
analysis however was indicative of diminishing average returns and not explicitly
marginal returns. Malthus tried to use the diminishing returns concept to make his
arguments regarding population growth, specifically, that the food supply was limited to
arithmetic growth. The concept remained part of economic thinking of the time, even
incorporated into the “four propositions” stated by Senior as he began the study of
political economy.
It was not until the twentieth century drew near that the distinctions were made
between average and marginal products. Clark presented a paper that applied the idea of
diminishing returns to all factors of production. His major assumption was that all
factors of production remain permanent except for one factor that would be changed.
Under these assumptions, if more units of the one varying factor were added, the
marginal and average products associated with that factor would eventually decrease.
Edgeworth in 1911 assumed that land was a fixed input and created a table that
included variable levels (referred to as “doses” of the other inputs and their resulting
output. Though he was arguing that these concepts would apply to any industry (in this
16
case, railways), his work was based on agricultural examples. Citing his observation that
at some point there is a transition from increasing to diminishing returns, Edgeworth was
one of the first recognize the marginal product of the input as well as the average product,
creating columns for each in his table.
Modeling Production
Thompson 1988 describes the flexible functional forms (FFF) as a way to relax the
restrictions one gets when using Cobb-Douglas functions. They can be expressed as
quadratics, Box-Cox, and numerous other forms. It is important that empirical studies
state clearly the reasons a particular form was chosen.
FFF are very useful when using duality theory. If the function satisfies certain
conditions (convexity, monotonicity, homogeneity), there is no longer a need for a self-
dual function. It is also possible to derive supply or demand from these functions and to
use them for comparative statics. Additionally, these functions can be used to estimate
equations (or systems) that are nonlinear in their parameters.
One of the main problems with the FFF is collinearity. Also, it is often difficult to
meet the above conditions over the entire set of observations. Estimating nonlinear
systems also limits interpretation, as much statistical theory assumes linearity in the
parameters.
Deiwert (1973) defines flexibility as a local property. Using an arbitrary function,
he makes a second order approximation. The parameters of the FFF then must give it
first and second derivatives equal to those of the arbitrary function. This definition of
flexibility is often applied because the conditions are easier to meet than those of other
definitions.
17
Another measure of flexibility is the Sobolev norm, which is a global definition.
This approach measures the average error of approximation and consequently estimates
elasticities very close to the true elasticity. This also gives the model nonparametric
properties including “small average bias approximations (Gallant 1981); consistent
estimations of substitution elasticities (Elbadawi, Gallant, Souza 1983); and
asymptotically size α testing procedures (Gallant 1982).”
Many FFF can be looked at as derivations from mathematical expansions, but the
definitions do not limit them to only such derivations. Some commonly used expansions
are the Taylor, Laurent, and Fourier expansions. The first two are flexible under the
Diewert definition, while the Fourier follows the Sobolev definition. Some functions,
like the generalized Mc Fadden and Barnett functions are not derived from an expansion
(Diewert and Wales 1987).
In Thompson’s analysis, four types of studies were used to look at the various FFF:
Monte Carlo, parametric modeling, Bayesian Analysis, and nonnested hypothesis testing.
The FFF are useful in both production and consumer applications of the Monte Carlo
studies, but depend upon the type of data, the size of the sample, and other properties that
vary. The parametric model used Box-Cox testing and therefore could only be applied to
some of the FFF. The Bayesian analysis allowed very different models to be compared
based on their data on both the production and consumer levels. The nonnested testing
includes all the proposed models and is also based on the data. Of these methods, the
Bayesian analysis and nonnested testing were the most useful in comparing FFF. It is
noted that in either case, it is important to be able to compare models with various
transformations of the dependent variable.
18
One must look at the duality of the behavioral model and test the behavioral
assumptions to make sure the data are consistent with the assumptions. The data should
then be tested for theoretical properties such as returns to scale. The chosen form should
fit the assumptions and can be compared with other forms through Bayesian analysis or
nonnested hypothesis testing. After the form has been chosen, it is important that it be
compared to other measures (in order to gauge sensitivity of that form).
In an examination of The Cottonwood River Watershed, Apland, Grainger, and
Strock (2004) describe a framework for creating a farm model that accounts for both
agricultural production as well as water quality concerns. In trying to model these
tradeoffs, Apland notes that mathematical programming models can incorporate
economic and agricultural factors but become very complex.
A deterministic farm model forms the basis of Apland’s work, which is designed
to be expanded to include risk. The model is made up of 18 production periods for the
year in order to represent harvest and planting activities in all combinations. Land, labor,
and machines are held fixed so that the analysis can focus on the various harvesting,
planting, and tillage dates and fertilizer application is endogenous.
To carry the model forward, the authors discuss a discrete stochastic
programming model (DSP). This type of model is useful in that it can capture alternative
practices as part of the risk analysis. Further, risk can be included in the constraints and
as par of the sequential decision process. However, this method requires a great deal of
data to be useful and can be costly.
Risk is a significant factor in modeling agricultural production. Just and Pope
(1979) note that risk affected by price, market phenomena, technology, and policy.
19
Traditional evaluations of production were drawn from experimental data, but the authors
argue that using continuous response functions give better estimates. This analysis uses
“neoclassical log-linear production functions”.
Some of the specific problems with previously used (and popular) models are that
increasing input always has positive marginal effect on output and that it also reduces
variability of marginal productivity. In order to separate the effects of input on output
and variability, the authors propose that a good stochastic specification has two functions;
the first modeling the effects of input on mean output and the second modeling the effects
of input on variance of output:
y = f(X) + h1/2(X)0, E(0) = 0, V(0)=1
The mean and variance of output can then been seen independently as E(y) = f(X)
and V(y) = h(X).
The procedure proposed for such and estimation is a three-step regression. The
first is a nonlinear least squares (NLS) regression of yield. Second, the expected error is
regressed against X using ordinary least squares (OLS). The final step is a weighted NLS
regression of y.
In using experimental data, the authors focus on Cobb-Douglas and translog
functions. The basic equation is then modified to include an error term, time, and plot to
capture time effects. They conclude that the simple two-part production function remains
within the bounds of traditional economic thought while removing some of the
constraints that hinder decision-making.
20
Modeling Production and Cost
In a static situation, Paris and Caputo (2004) state, any estimation of the economic
relationships of a price-taking firm should include both primal and dual relations. Their
proposed model uses a generalized additive error (GAE) approach to make a nonlinear
estimation. Their sample firm is at once risk-neutral and cost minimizing. The system is
then represented as a set of equations: the primal production and input price functions,
and the dual input demand function.
This analysis does face some challenges. The planning and decision-making data
is generally not recorded by producers and thus has to be estimated. Also, the choice of
production function can pose additional challenges. The Cobb-Douglas and constant
elasticity of substitution approaches have the same functional form as their respective
cost functions, but that may not be the case with other forms. Once quantities and prices
are estimated using NLS, they are put into a nonlinear seemingly unrelated (NSUR)
model.
Economics of Water Use
Water use, an important factor in production, is often modeled as water applied.
However, Kim and Schaible (2000) challenge the assumption that water (or any of the
variable factors) is completely engaged in crop production. Noting that the production
process does not consume not all inputs applied, whether water or fertilizer, the authors
seek to provide a measure of the overestimation of economic benefits from water use.
The authors observe that economic benefits in agriculture are often modeled as
normalized-quadratic functions, but that the derived factor demands are sometimes linear
and sometimes in Cobb-Douglas form. As such, both linear and nonlinear cases are
examined. Under both scenarios, the total economic benefits were overestimated when
21
using applied water rather than consumptive water. In the application of these methods
to corn production in three Nebraska counties, the overestimation was 28.9% and
estimate even higher when looking at agriculture overall.
In one of the most traditional models of water use, the Von Liebig production
function as described by Boggess et al (1993), uses evaporation and transpiration to
estimate the changes in yield. This type of function describes a linear output relationship
until some maximum. From that equation, actual yield can then be estimated as a
function of the ratio of actual to potential evapotraspiration.
Similar to Kim and Schaible, Boggess et al make the distinction between effective
water, actually used by plants, and the total amount of applied water. In modeling
irrigation, they point out that it is fundamental to incorporate the concept of hydrologic
balance. The principle of hydrologic balance states that there must be equality between
the amount of water that enters a specific area and the amount of water that leaves that
area. That is, all water entering the particular area, through precipitation, irrigation, or
from the soil, and all water leaving the area whether through evapotranspiration, runoff,
or percolation must be considered.
South Florida Agriculture and Ecosystem Restoration
Restoring the water flow of the Everglades will create a need for water retention in
the northern part of the watershed. By 1978, over three million acres of land had been
drained in South Florida (Weisskoff 2005). This region, especially the Everglades
Agricultural Area (EAA) will require a great deal of water to meet needs during the dry
season. Development of the EAA created a system if irrigation and drainage that
prevents most water retention during the wet season. Increasing the amount of water
retained may lessen agricultural profitability. Authors Aillery, Shoemaker, and Caswell
22
attempt to model the economic effects of water table management and surface water
retention scenarios.
The authors measured the tradeoffs under two conditions, the first being that
resource use is determined by agricultural producers alone, the second using joint
maximization of agricultural and environmental objectives. Using both objectives
resulted in higher marginal costs, but also significantly increased the benefits of lowering
the water table (Aillery, et al 2001). Whether the benefits will outweigh the cost is
dependent upon the specific agricultural and environmental demands.
Three scenarios of water policy were simulated: water-table restrictions, surface-
water development (including land acquisition), and water-retention targets. The first
showed an increase in water retention, but at a high opportunity cost and the inability of
the soils to retain the desired amount of water. Surface water development also increased
water retention, but comes at the cost of production foregone by retiring those lands. A
moderate change to a target water-table depth was considered the best option (Aillery, et
al 2001).
The authors are up-front about two main concerns with this article. The first is
that a true cost-benefit analysis would need more empirical evidence from the agricultural
sector and is difficult to generalize. The second is that sugar prices, the major component
in estimating agricultural gains reflect price support levels that could change in the future
and thus alter these findings.
Sugarcane response to high water tables and flooding
Glaz, et al (2002) note that the EAA is dependent upon the canal system to
maintain suitable water levels for sugarcane and other crops. Pointing to a study by
Omary and Izuno (1995), ideal water levels fall within 40 and 95 cm below the soil
23
surface. Keeping the water within this range has become more difficult as the farmers are
dealing with soil subsidence. As soil is lost, the remaining soil cannot store as much
water. Additionally, best management practices put in place to limit Phosphorus entering
the Everglades have resulted in farmers pumping less water and thus maintaining higher
water levels.
The researchers conducted two experiments, the first being planted in February and
harvested in three cycles (plant cane and two ratoon crops). The second experiment was
planted in January of the following year and harvested in two cycles. In both cases, two
fields were planted and received water treatments from June to October (the months with
highest rainfall). The wetter field was treated to have a water level between flooding and
15 cm BSS. The drier field was kept with a water level between 15 and 38cm BSS.
Over the period of study, the researchers found that the soil profile was very
sensitive to rainfall. They estimated that for every cm of rain, the soil profile rose 10cm.
As a result, even the drier field had some days in which the water level was higher than
15cm BSS, suggesting that during a normal year, fields with the drier target would still
see flooding.
For the plant cane crops, the average cane yield for the drier field was 5.8% higher
than the wetter field. For the first-ratoon harvests, the drier field average cane yield was
4.3% higher than the wetter field. For the second-ratoon harvest, the average cane yield
was 8.4% higher in the drier field.
The researchers also measured the sugar yields from each harvest. The average
sugar yield for the plant cane crops was 6.6% higher in the drier field than in the wetter
field. The average sugar yield in the first-ratoon crops was 8.3% higher then in the wetter
24
field. In the second-ratoon crop, the average sugar yield was 11.5% higher than in the
wetter field.
The project suggests that there are some cultivars that may continue to yield well if
the water tables were maintained at a higher level. They suggest that increasing the water
table incrementally would be the best option considering profit and the need to reduce
phosphorus discharge.
In 2004, Glaz et al published their findings after experimenting with two different
sugarcane genotypes. This study was to examine periodic flooding, lasting no longer
than one week and then draining to a water depth of about 50 cm below the surface.
Flooding was set at 7 days to simulate the longest flood period a commercial field in the
EAA might experience. The areas were treated for five and nine cycles in different years.
It was noted that in the EAA, it is often difficult to drain to the desired level after a flood.
One of the genotypes developed arenchyma (air cavities) at the roots, which seems
to have impacted the yield response. These did not show a significant response to
changes in depth over the three years. The other genotype however, showed a 21% cane
yield increase and an 18% sugar yield increase in the fully drained case as compared to
the flooded specimens in 2000. The 2002 experiment resulted in a 28% increase in both
sugar and cane yields in the drained area compared to the flooded plants. The authors
point out that using additional flooding periods might result in a nonlinear flood response.
Such information would be of use to farmers that are not able to drain all fields at once
due to limitations on total drainage to the canals.
Another consideration in sugarcane response is the possible benefit of flooding
during certain stages of growth, specifically in trying to control for wireworm (Glaz
25
2002, 2003). Flooding the seed cane after planting could replace the practice of applying
insecticide to the soil. Before the practice could be commercially adopted, however, the
feasibility and cost of maintaining the flood condition and then draining would need
further examination. There is also a concern that the shortening of the growing season
would lead to reduced yield.
26
CHAPTER 3 METHODOLOGY
Introduction and Overview of Analysis
The analysis assumed a hypothetical 640-acre sugarcane farm. Using two different
approaches, one based on agronomic yield models and historical groundwater levels; the
other based on historical yield and rainfall, this analysis estimated profits foregone for
this “typical” farm. Both approaches assume that the operation is independent of a
processor and profit maximizing.
In the first approach, the agronomic functions measuring response to flooding/high
water tables were combined to give an expected yield function that included a parameter
for flood events. That function, along with acreage, overhead cost, variable cost, and
price were then simulated, varying the input factors. This approach also used a historical
distribution along with a range of likely water table levels as described in Water
Management for Florida Sugarcane Production (2002) and Agriculture and Ecosystem
Restoration in South Florida: Assessing Trade-offs from Water-Retention Development in
the Everglades Agricultural Area (2001).
The second approach utilized historical yield and rainfall, determining relationship
between the two based on the most sensitive growth period. Future rainfall will be based
on historical records and used to provide possible yields. From the previous Water
Management article, the EAA Storage Reservoir Phase 1 Existing Flood Control
Conditions Documentation, and the 2001 study on drainage uniformity, this approach
will assume that the system will drain up to 48% of rainfall.
27
Agronomic Model
Taking the findings of the empirical research on yield response by Glaz et al
(2002), two equations were combined in order to incorporate a parameter for flood
events. The results for the two experiments were as follows:
Y x= +14 6 016. . (1) Y x= +17 6 25. . (2)
The year 2000 experiment with 5 flood cycles (Eq. 1) was defined as a base (Z=0)
and the 2001 experiment with 9 flood cycles (Eq. 2) was defined as Z=1. The following,
then, represents flood events, Z, as
Z = -1.25 + .25F,
where F is the number of flood cycles, and Z is a qualitative variable. The number of
flood cycles was simulated as part of the analysis. The simulation of flood cycles was
first attempted using a uniform distribution (between 0 and 9), but ultimately, F was
determined using a triangular distribution from which pseudo-random numbers were
generated. The boundaries of the distribution remained 0 to 9 in keeping with the
experimental conditions.
Equations (1) and (2), can then be combined as
Y = 14.6 + 3(Z) + .16X +.09(Z)X.
Where Y is yield in kilograms per meter squared. Sugarcane yield is historically
measured in tons per acre, so the resulting yield (in kg) must be converted by a factor of
approximately 4.5 to be expressed in tons per acre (see Table 3-1). In order to translate
the empirical data to practical terms, a calibration factor was also included. This factor
(0.267) was determined by setting the mean value of the empirical yield data equal to the
historical mean yield.
28
The other key variable in this model is the groundwater level. Table 1 illustrates all
key output variables (KOV) and their distributions. In order to simulate the probable
range of groundwater levels, the historical data determine the distribution from which the
simulated values will be chosen. A normal probability plot suggested that the data were
very close to a normal distribution. The ANOVA procedure was used to determine the
mean and standard deviation for the groundwater variable based on these data. The mean
would be adjusted in order to simulate various scenarios.
Table 3-1. Key output variables and probability distributions for empirical model Key Output Variables Specifications Acres 640 Flood Events "Z" Z = -1.25 + .25F Flood Cycles "F" Triangular distribution: 0 to 9, Depth “X” cm Varied by scenario Yield (kg per meter sq.) Y = 14.6 + 3(Z) + .16X +.09(Z)X Yield (tons per acre) Y * 4.460947/3.74 Price (per ton) $31.70 Variable Cost (per acre) $760.94 Total Fixed Cost (dollars) $144,000 Profit (Price* Yield)-Total Costs
. Using the Simetar (Richardson, 2001) simulation tool, these data were simulated
for 100 iterations for each scenario. The output for each could then be compared in order
to determine the effects of new water conditions. Five different scenarios were simulated
using this tool.
• Scenario 1: A representation of current conditions, this scenario assumed a mean
water table depth of 85.2cm and a standard deviation of 43.23 based on USGS data. • Scenario 2: Also represents current conditions, but with the mean depth adjusted to
76.2cm as described in Water Management for Florida Sugarcane Production • Scenario 3:A model of post-restoration conditions by raising water table depth to
54.78cm as suggested by Aillery, Shoemaker, and Caswell (Scenario I-5, 2001). • Scenario 4: Alternative model of post-restoration including a truncated normal
distribution for water table depth with mean 50cm, minimum -27.1272cm
29
(historical minimum), and maximum 92.8cm (95% Upper Confidence Interval for historical data).
• Scenario 5: Identifies conditions under which the hypothetical operation exhibits zero profit.
Rainfall Model
This approach began with the ANOVA procedure on monthly rainfall data over a
twenty-year period in order to find variation for further simulation. Also included were
the values for pan evaporation (Evap) and average temperature (Temp) for each month
over the same growing period. The data were aggregated such that the annual yields
were matched with the previous growing period. For example, Rainfall summed from
August of 1990 to January 1991 corresponded to the 1991 production data. Maximum
drainage (Drain) was set at 48% of the rainfall for each of the months included.
A relationship between total production (TP) and these factors was determined by
using an Ordinary Least Squares (OLS) regression:
TP = 2464440.7 - 2947.3Evap - 299037.5AugRn - 333598.7SeptRn - 266891OctRn - 303408.3NovRn - 309797.5DecRn - 258229.4JanRn -7455.9Temp+ 645862.7Drain
Using this relationship, normal distributions for rainfall were specified for the
simulation based on historical mean and variance
In order to represent the changes in drainage practices, the post-restoration scenario
changes the percentage of water drained was varied while other climatic factors were
held constant. The results of the two scenarios could then be compared to each other and
ultimately to the previous model.
The rainfall model was designed to capture the concept of waster balance as
presented in the literature regarding water use. It was anticipated that that in defining the
amount of water coming into or out of the given system, future water flows could be
30
estimated and consequently changes in production could be predicted. In this case,
varying the drainage capacity could give distinct scenarios for comparison.
31
CHAPTER 4 DATA SOURCES
Empirical Research on Sugarcane Response
Glaz et al (2004) examined water table effects on two sugarcane genotypes in
experiments from 2000-2002. Previous research, they noted, was inconsistent regarding
sugarcane response to water tables. EAA farmers specifically have to deal with periodic
floods (less than a week) and cannot always drain the desired amount of water. To
simulate these conditions, the experiment evaluated periodic flooding followed by
drainage to depths of 50, 33, and 16 cm below the soil surface (BSS).
To carry out the study, lysimeters made of polyethylene (1.5m x 2.6m x 0.6m) were
set up with Pahokee muck soil from an uncropped EAA field. Each lysimeter had well
water flowing in each day and a pump to get rid of excess water. Additionally, each had
a valve that drained the lysimeter to the target water table level. Two sugarcane
genotypes were planted, both being chosen because of high yield and similarity to
commercially produced varieties. After planting, the water level remained at 50cm BSS
until the actual treatments started. There were four total treatments; one a control and the
others being flooded for the first week of a three-week cycle. After the 7 days of
flooding, these treatments were drained to the aforementioned depths.
Water height from the actual soil surface up to 2.5cm above the surface was
considered a “flood” condition in this experiment. The length of the flood, 7 days, was
set to simulate the longest period of flooding one could expect in the EAA. Similarly, the
50cm control depth was based commercial practices. The experiment was repeated for
32
three years; the 2000 experiment used five flooding cycles while the 2001 and 2002
experiments used nine flooding cycles. The resulting response functions were
Y x= +14 6 016. . (2000, r 0 992 = . )
Y x= +17 6 25. . (2001, r 0 942 = . )
Where Y is equal to cane yield in kg m-2 and x is equal to water table depth (cm)
during drainage. The difference in flood cycles can then be used as a factor in
incorporating the number of floods into the yield response analysis.
Water Table and Flooding Conditions
To be consistent with the cultural practices of the EAA, establishing the possible
range of water tables included water table levels as described in Lang et al’s Water
Management for Florida Sugarcane Production (2002). They noted that 30 inches
(76.2cm) was optimal depth in terms of sugarcane yield and stated that the recommended
target level would be a depth of 23-30 inches (58.42-76.2cm) to the surface. They also
noted that variation in EREC studies ranged from 39 inches (99.06cm) to surface level.
Historical water table levels for the area were available from the USGS from
October of 1977 to September of 1995. The variation here ranged from a maximum
depth of 206.95 cm below the surface to a minimum of just over 27 cm above the surface.
The mean depth was just over 85 cm below the surface. The 167 observations, however,
are at irregular intervals, which made the information useful only in determining the
variation in water table levels. The complete distribution of these data is represented in
Figure 4-1. These data were collected from a well at Latitude 26°38'45", Longitude
81°26'07" in Hendry County, Florida.
33
PDF Approximation
-50.00 0.00 50.00 100.00 150.00 200.00 250.00
CM below surface
Figure 4-1. Distribution of water level for Hendry County, FL 1977-1995. Source: USGS
Additionally, a more qualitative source of information on water management was a
summary of meetings with EAA sugar growers in November 2003 to get a consensus on
flooding conditions. Coordinated by the Southwest Florida Water Management District,
the EAA Storage Reservoir Phase 1 Existing Flood Control Conditions Documentation,
provided insight into the growers’ major concerns. The participating groups included US
Sugar, the Sugar Farms Cooperative, and Florida Crystals, all producers within the EAA.
The documentation of the three meetings revealed a number of common points. Some of
these key statements included:
• Farmers have not kept regular records of crop losses due to flooding thus far • The sugarcane crop is most sensitive to flooding during early stages of growth • Receiving more than 4 inches of rainfall in a 24-hour period is considered
problematic There was also consensus among the growers that heavy rainfall and flooding are
of most concern to the areas near the Bolles and Cross Canals which provide water flows
to the east and west of the EAA. There was a concern that these canals to no have the
capacity to carry water out as needed.
From the 2001 study on drainage uniformity it was noted that sites normally
drained an average of only 48% of the rainfall input into the system (Garcia, Izuno,
Scarlatos). When looking at the flow of water in and out of the EAA farm system, this
34
average will used to determine how much water is being drained out of the system rather
than contributing to the groundwater level.
Historical Production
Through the National Agricultural Statistics Service (NASS), the US Department
of Agriculture (USDA) provides historical production information down to a county
level. Using the “Quick Stats” website allows users to search production history for all
major crops. Under the category Sugarcane for Sugar, county-level data are available for
acres harvested, yield per acre and total production. Figure 4-2 illustrates the total annual
production for the Hendry County over the past ten years.
0
0.5
1
1.5
2
2.5
3
Mill
ions
of t
ons
1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004
Annual Sugarcane Production in Hendry County
Figure 4-2: Total sugarcane production for Hendry County, FL from 1994-2004. Source: NASS
The production values are listed annually and are available from 1977-2004. For
this analysis, however, data from 1980-2004 were considered. The area harvested over
that time ranged from 35,000-76,000 acres. The average yield per acre varied from a low
of 28.6 tons in 1981 to a maximum of 40.1 tons in 1998. Total annual sugarcane
production in the county peaked in 2002 at over 25.8 million tons.
35
Climatic Data
The Everglades Research and Education Center includes an automated weather
station that provides climatological data as a cooperative project of the University of
Florida / IFAS and the South Florida Water Management District. The station was
provides data on temperature, rainfall, solar radiation, and evaporation at the coordinates
26.6567N and 80.6299W. The time series for rainfall goes back as far as 1924, but for the
purposes of this analysis, only data from 1979 – 2000 were considered. The total
evaporation and average temperature over the growing season were also noted.
Within the area of interest, Hendry County, FL, sugarcane planting takes place
from August through January and this period is considered the most sensitive to excess
water. Table 4-1 illustrates the total rainfall for those crucial months.
Table 4-1. Total monthly rainfall in inches for the EAA 1979-2000 . Year January August September October NovemberDecember1979 5.39 14.24 2.24 4.80 2.89 1980 6.09 3.73 7.46 1.02 3.42 0.73 1981 0.68 17.37 4.87 0.67 3.08 0.85 1982 0.63 8.11 12.41 2.75 0.67 0.80 1983 3.91 6.26 6.77 5.16 1.16 4.42 1984 0.23 4.04 8.15 0.40 2.37 0.11 1985 0.75 5.52 9.63 3.39 1.54 2.20 1986 3.59 6.21 4.04 4.50 1.58 3.99 1987 1.18 4.20 4.49 3.14 8.04 0.30 1988 3.02 8.89 2.47 0.11 1.31 0.89 1989 0.97 6.92 8.91 3.49 1.24 1.95 1990 2.47 7.57 2.96 4.22 0.39 1.11 1991 8.24 2.83 6.27 3.54 2.45 0.46 1992 1.20 11.85 10.82 0.69 4.03 0.62 1993 10.16 6.19 5.56 8.00 1.75 0.79 1994 5.60 9.74 5.47 12.16 5.93 7.13 1995 1.91 10.51 8.76 9.60 0.65 0.89 1996 1.35 9.75 3.04 4.23 0.80 0.38 1997 1.23 4.56 5.47 0.65 4.44 5.77 1998 1.47 9.37 11.64 2.20 11.25 1.00 1999 1.95 5.04 8.18 7.69 0.72 0.45 2000 0.74 3.58 7.00 4.77 0.54 0.24
36
The normal distributions for rainfall during these months, specified through OLS
regression, were used to in a simulation to predict the possible rainfall in each of the
crucial months (Table 4-2).
Table 4-2. Average EAA rainfall.
Month
Mean Rainfall (inches)
Standard Deviation
January 2.73 2.69 August 7.25 3.45 September 6.87 2.84 October 3.92 3.21 November 2.73 2.77 December 1.67 1.96
Source: EREC Weather
Costs of Production
Cost and Returns for Sugarcane on Muck Soils in Florida (Alvarez and
Schuneman) provides a framework for looking at the production costs for farming
sugarcane in the EAA. This work provides a number of key assumptions including: a
profit maximizing management, independent grower status (non-producer), a small farm
unit, and a three-crop cycle, that is, the hypothetical farm crop includes first planting and
first and second ratoon crops.
However, cultivation and harvesting practices have changed and the most recent
data regarding production costs comes from the USDA Economic Research Service
(ERS). These data (see Table 4-2) from the Farm Business Economics Report take into
account the additional Everglades Restoration tax that began in 1995. The 1995-1996
values were the final values published in this form. For the purposes of this analysis, it
will be assumed that all acreage in the model is harvested.
37
Table 4-3. Florida sugarcane production expenses
Item Dollars per Harvested Acre
1995 1996 Variable Expenses Seed 27.95 27.95 Fertilizer 61.33 57.99 Chemicals 59.10 61.15 Custom operations 106.65 104.81 Fuel, lube, electric 21.67 23.79 Repairs 80.12 80.84 Labor 406.54 396.76 Irrigation water purchased 6.70 7.07 Miscellanous 0.56 0.58 Total Variable Expenses 770.62 760.94 Fixed Expenses General farm overhead 114.79 107.45 Taxes and insurance 59.27 59.93 Interest 9.61 9.49 Total Fixed Expenses 183.67 176.87 Total expenses 954.29 937.81
Source: USDA, ERS
Sugarcane Prices
The Florida Agricultural Statistics Service (FASS) maintains an annual record of
acreage, yield, production, season average price and the overall value of production.
From these field crop summaries, prices were recorded from 1980-2003. These prices
reflect sales of sugarcane for sugar and seed. As illustrated in Figure 4-3, there has not
been a great deal of volatility in price. The maximum price, $39.40 per ton, occurred in
1980 and was followed by a 27% drop to 28.60. After that initial fall, prices have
remained close to $30 a ton. In contrast, the lowest season average price was $27.20 in
1999.
38
Annual Price of Florida Sugarcane
0.005.00
10.0015.0020.0025.0030.0035.0040.0045.00
1980 1985 1990 1995 2000
Dol
lars
season avg price
Figure 4-3: Season average price: sugarcane for sugar and seed 1980-2003. Source: FASS
It is likely that a major factor in maintaining this price stability is the tariff-rate
quota system in place on sugar imports. This analysis, however assumes that these
conditions will not change and thus do not factor into the modeled scenarios.
39
CHAPTER 5 RESULTS AND DISCUSSION
The analysis involved creating two stochastic models based on agronomic studies
and historical data regarding sugarcane production and growing conditions in the
Everglades Agricultural Area. Once each model was specified, the key output variables
were identified and their respective probability distributions defined. The Simetar
simulation application generated pseudo-random numbers, based on the given probability
distributions, which were then used to update the model equations 100 times. Each
iteration of the simulation generated new values for the every key variable. However, the
values for water table depth, yield, and profit (on the basis of hypothetical 640-acre farm)
selected as the output of the simulation, as these were of most interest.
Empirical Model Results
Five scenarios using an empirical yield model were completed in order to compare
production possibilities with and without the restoration conditions, emphasizing
maintenance of a higher water table. These included:
• Scenario 1: Current conditions, assuming mean water table depth and standard deviation based on USGS data.
• Scenario 2: Current conditions, but with the mean depth adjusted to 76.2cm as recommended in Sugarcane Production literature.
• Scenario 3: Post-restoration conditions with water table depth to 54.78cm as suggested by Aillery, Shoemaker, and Caswell (Scenario I-5, 2001).
• Scenario 4: Post-restoration incorporating a truncated normal distribution for water table depth.
• Scenario 5: Zero-profit condition The complete outputs for these simulations are illustrated in Appendices A-D. A
summary of the simulated profits is illustrated in Table 5-1, recalling that the fifth
40
scenario was the zero-profit condition. The output series were also compared in pairs in
order to determine whether the results were statistically different across the scenarios.
Table 5-1. Summary statistics for simulated model output, profit in dollars* Scenario 1 Scenario 2 Scenario 3 Scenario 4
Mean 26,722.33 (5,370.24) (85,485.58) (135,442.80)Standard Deviation 215,148.12 209,707.29 199,273.51 142647.91 Minimum Value (378,281.63) (388,859.30) (550,047.54) (383,743.63)Maximum Value 656,710.50 642,406.47 656,041.87 239,375.77
* For 640 acres of production In terms of the simulated profits, it was surprising that the mean value for Scenario
2, which incorporated the current conditions, recommended in the Sugarcane Production
literature. In all other respects, the two base scenarios (1 and 2) were expectedly similar.
There is still a large range in the output, potential losses of over $500,000 to profits of
over $640,000, but the truncated distribution of Scenario 4 appears to have been most
successful in reducing the extreme values. It must be noted, however, that even that
scenario exhibits more variation than should be expected.
As the model stands, economic losses are probable. Scenario 5 was indeed a
zero-profit scenario, in which the original specifications were solved to determine the
point at which total revenue equaled total cost for the sugarcane operation. The zero-
profit conditions were: water table depth of 31.83cm, 6 flood cycles (for a Z value of .24),
resulting in a yield of 31.1 tons per acre, just a 7 % difference from the historical mean.
The mean yield for Scenario 1, 32.42 tons per acre as stated in Table 5-2, is indeed
comparable to the historical average of 33.48 tons per acre produced in Hendry County
from 1980-2001. The post–restoration scenarios showed a substantial drop in yield to
approximately 27 and 24 tons per acre. However, the variation within this model is still
greater than one sees across the historical data. Specifically, the standard deviation is not
consistent with the historical standard deviation of 2.67
41
Table 5-2. Summary statistics for simulated model output, yield in tons per acre* Scenario 1 Scenario 2 Scenario 3 Scenario 4 Mean 32.42 30.84 26.89 24.43 Standard Deviation 10.60 10.34 9.82 7.03 Minimum Value 12.46 11.94 3.99 12.19 Maximum Value 63.47 62.77 63.44 42.90
*Rounded to two decimal places .
Regarding the modeling for water table depth, the historical mean and distribution
were successful in providing model results that were reasonable considering both the
historical information as well as the range suggested in the water management guidelines
and literature. In this instance it seems it was successful to maintain variability while
adjusting the means.
Rainfall Model Results
As previously discussed, multiple regression using the historical data on
production, rainfall during crucial months, evaporation, temperature, along with likely
drainage levels resulted in the following relationship between total production (TP) and
the climatic factors:
TP = 2464440.7 - 2947.3Evap - 299037.5AugRn - 333598.7SeptRn - 266891OctRn - 303408.3NovRn - 309797.5DecRn - 258229.4JanRn -7455.9Temp+ 645862.7Drain
The simulation was to provide outputs including yield per acre, total production,
and profit for the 640-acre farm. Analysis of this model, however, revealed that it lacked
the explanatory power necessary to be of use in decision-making. While it was not
unexpected that the historical data would have a great deal of error, selected variables
explained only 48% of the variation in total production from 1980-2001. When
regressing the same variables against the historical yields, the results explained even less
42
variation, with an R2 value of approximately .26. Though the historical model was not
suitable for this analysis as specified, it may be useful in decision-making if modified.
Comparison of Model Scenarios
While the current and post-restoration scenarios remained linear models, the
constraint on the distribution of water levels changed the slope of the trend when
comparing water levels with profit for the hypothetical sugarcane farm. The range of
simulated values was wider than historically expected across all the scenarios. The
change in water table levels, constraining the possible distribution, also had a great
impact on the possible production range simulated. The variation in scenarios 3 and 4
was predictably more limited across all variables.
Only Scenario 1 yielded a positive average profit once simulated. It was more
surprising, however that the other current scenario, using recommended optimal depth
would average in the negative. Compared to the zero-profit case, Scenarios 3, 4, and 5
produced a mean water table that was higher than the zero-profit solution as would be
expected.
There remained a great deal of variation within the simulation results across all
cases. Even Scenario 4, designed to reign in some of this variation by using a truncated
distribution produced results ranging from losses of over $380,000 to profits of nearly
$240,000. Tracing the variation back to the simulated yield, one must note that Scenarios
1, 2, and 3 resulted in maxima that are far beyond current or historical yield levels. On
the other end of the spectrum, Scenario 3 produced a minimum that was similarly
improbable, a mere four tons per acre.
43
Evaluation of Hypotheses
The two hypotheses for this study were analyzed to determine whether there was a
significant difference in the mean or variance using a two-sample t-test and an F-test.
The tests compared the output series of the simulations from both scenarios of the
empirical model. The following hypotheses were tested:
Maintaining a Higher Water Table Lowers Average Sugarcane Production for an EAA Farm.
A distribution comparison of yield results indicated that both the mean and (and in
most cases) variance of the base model and the post-restoration scenarios are statistically
different. Given a confidence level of 95%, the null hypothesis can be rejected. The two
base scenarios, however, were not statistically different from each other in terms of the
production results. Additionally, the variances of each base scenario compared with
Scenario 3 were not statistically different. Though the difference between scenarios was
significant, it is essential to note that the post-restoration scenario did provide some
yields at or above the current production levels.
The basic functions used at the beginning of the analysis indicated that yield would
be responsive to water changes. Once modified to represent an EAA farm, this model
indicated that a shift in the range of maintained water levels affects the possible yields for
that farm. The rejection of the null hypothesis in this case indicated that for this type of
farm, the adjustment to higher water tables results in lower yield on average. As there is
still variation in those water levels, the possibility exists that the farm could achieve
greater yields. However, it is more likely that the sugarcane farm, under these conditions,
will see lower yields.
44
The EAA Sugarcane Operation Will Experience a Reduction in Profit Under the Changed Water Conditions.
In examining the assumed sugarcane operation, analysis of the simulation series
for profit indicated that these results were also statistically different for both mean and
variance across most scenarios. The null hypothesis can then be rejected with a
confidence level of 95% except in comparing the mean results between the two base
scenarios; that is we can state that there is a significant difference in expected profit
between each base scenario and the post-restoration scenarios. It should also be noted that
between scenarios 1 and 3 and also between 1 and 4, the F-tests are such that the null
hypothesis for variance cannot be rejected.
As in the previous case, though the series are statistically different, they were not
mutually exclusive. Of particular importance regarding the difference in profit is that the
given simulations kept harvested acreage constant, a variable that directly impacts the
costs of production. Similar to the yield hypothesis, it is likely that the operation will
experience losses, but not absolutely certain.
Looking at the assumed EAA sugar operation, failure to reject the null hypothesis
demonstrates that this “typical” operation will indeed see changes in profit in response to
varying water conditions.
45
CHAPTER 6 SUMMARY AND CONCLUSIONS
Summary
The purpose of this research was to provide an economic perspective in examining
the impacts of the Comprehensive Everglades Restoration Plan within in the Everglades
Agricultural Area. As sugarcane is the dominant crop in the area, a farm-level analysis
was developed in order to examine the impacts of water regime changes on sugarcane
production. For the purposes of this analysis, those changes were limited in scope;
defined as flood events and a higher average water table. In keeping with previous
authors’ work on cots of production, the hypothetical farm was modeled assuming that
640 acres in production, the firm employs profit maximizing management, and
independent grower status.
The analysis incorporated an empirical water response function and simulations of
five scenarios represent the possible water regimes. In order to better align with
historical sugarcane production, the empirical data were calibrated. Additionally, the
model incorporated historical cost and water distributions based on USGS records. A
model based on historical yields and weather data was also developed, but lacked
explanatory power and consequently was not used in a simulation. Only one of the
modeled scenarios showed an average profit over the course of the simulation.
Conclusions
There was a significant difference between the simulated bases and the post-
restoration models, confirmed by a two-sample t-test and an F-test. Upon examining the
46
mean values of these scenarios, one could expect an EAA sugarcane farm to incur
monetary losses. Though these results are specific to a small, independent farm, the
probable losses are important on a regional scale. These results illustrate the additional
costs of CERP implementation beyond construction and planning. This one farm
situation is representative of the tradeoffs farmers are facing in the EAA. Best
Management Practices help fight soil subsidence and excess nutrient runoff, which are
beneficial to all in the long term. However, these immediate losses are considerable.
If this hypothetical operation is indeed indicative of EAA farming, the sheer
volume of sugarcane production magnifies the impact. It is important to remember that
sugarcane is the primary crop in the EAA, making up 86% of the area’s acreage in
agricultural production. The EAA generates nearly $800 million in producing nearly
25% of domestic sugar. While this analysis did not attempt to specify total regional
implications, it stands to reason that the entire area would be impacted. It is highly
unlikely that all farms will experience the same degree of loss, but the high value of this
crop and its regional importance suggests that there are many stakeholders who would be
adversely affected.
The results of the simulations were not completely negative, however. Even in
the constrained post-restoration model, there were profitable iterations. That is, even
under changed conditions, part of the range of conditions is such that yields would
remain high enough to be profitable. In practical applications, the profit-maximizing
producer will do everything in his power to remain in that profitable range. It is also
important to remember that while many conditions in the model were fixed, there is a
great deal of research going into new cultivars of sugarcane, some of which are being
47
bred for resistance to flooding and/or high water tables. The producer facing a scenario
as described would do well to examine the possibility of using such cultivars.
There were, however, a number of limitations on the study. One of the primary
limitations was a lack of available, uniform data. In some cases, such as in trying to
determine flood information, it was clear that producers did not generally keep records of
specific flood events or the resulting damage. Similarly, the analysis would have
benefited from having more specific historical yield information rather than relying on
and annual average. Additionally, the assumptions made in defining the study farm may
hinder the application of information gained. Few sugarcane operations are independent
of sugar processors and the division of first planting and consequent ratoon crops is most
likely not uniform as in this analysis.
The estimation of production costs is another limitation to the relevance of the
study findings. As many of the sugarcane farms are part of vertically integrated
production and processing operation, it is difficult to get accurate estimates of production
revenue. In order for this analysis to be useful to producers or water management
officials, it would be necessary to update the cost portion of the model. A new
publication on costs and returns will soon be available (J. Alvarez, personal
communication) and those would greatly improve the value of the analysis.
The complexity and scope of issues surrounding production in the EAA make it
difficult to incorporate all factors into a single analysis. For example, a more intricate
analysis would have to consider the introduction of flood resistant cultivars and the
interaction of price supports on domestic sugar. The entire restoration effort is a
multidisciplinary project and though the analysis incorporated a variety of perspectives
48
and sources of information, the problem calls for the attention and evaluation of experts
from many disciplines.
Implications for Future Analysis
It is intended that this research might be an initial inquiry that could lead to further
study by those in water management or perhaps even producers. As a first step, this
study sets forth a framework that identifies key variables, such as maintained water level,
which are essential in the planning process. By relaxing some of the assumptions and
updating certain data, this type of analysis could be easily replicated and used by those
most interested in making efficient water management or production decisions.
With updated cost and inputs, additional scenarios can be simulated using the same
model, which lends itself to repeated analysis throughout the decision-making process.
Water managers could benefit form the additional information when adjusting the
capacity for water storage or drainage. With improvements, even the climatic model
could be of use to producers looking to analyze the balance of water in the production
system, taking into account rainfall, evaporation, and drainage. Such a model would also
have to account for irrigation, which was not available for this particular analysis.
It had been hoped that this analysis could capture the current production practices
in the area, describing the crop response such that a producer would have knowledge of
how changes in those practices could affect his operation. However, the large range of
output is an indication that this particular analysis is not yet ready for field application.
Considering the economic importance of this industry, it follows that further analysis and
continued data collection are warranted.
Over the course of this study, it became evident that decision-makers in the EAA
have to make policy and production decisions without complete information. With
49
shared information, such as that from the referenced growers’ meeting on flooding, and
tools like spreadsheet simulation, future work should be able to better refine the problems
to be researched and provide relevant information to all stakeholders.
50
APPENDIX A SIMULATION OUTPUT FOR SCENARIO 1
Iteration "Z" "F" X(depth)Yld tons/ac Profit
1 -0.14087 4.43652 36.42929 22.99319 -164516 2 0.262596 6.050384 2.020689 18.53983 -254866 3 0.293122 6.172489 88.79114 37.68041 133458.6 4 0.809249 8.236997 99.84259 47.38158 330275.8 5 0.621738 7.486952 136.1192 53.95415 463620.3 6 -0.70749 2.170035 42.67497 19.51556 -235070 7 0.267268 6.069074 71.53518 33.60959 50869.81 8 0.136634 5.546537 24.53701 22.63242 -171835 9 0.003309 5.013236 84.00906 33.03105 39132.38 10 0.551755 7.207022 91.06506 41.58563 212687.7 11 0.098492 5.393969 59.76264 29.39677 -34600 12 -0.44181 3.232743 107.8567 30.87399 -4630.05 13 -0.49417 3.023337 56.77932 23.14934 -161348 14 -0.37266 3.509378 114.7882 32.93913 37267.45 15 -0.59678 2.612878 120.9387 30.19318 -18442.4 16 -0.77456 1.90174 5.295971 15.00526 -326575 17 0.028506 5.114025 38.84618 24.70655 -129755 18 -0.88057 1.477735 54.88547 19.28265 -239795 19 -0.62596 2.496178 82.94403 25.08287 -122120 20 -0.0305 4.877987 145.6574 44.01623 261999.8 21 -0.17373 4.305074 111.1101 35.43426 87888.72 22 -0.32012 3.719529 157.6971 40.38569 188343.3 23 -1.12519 0.499255 81.78061 18.8561 -248449 24 0.456235 6.824941 118.418 46.7976 318428.2 25 0.194742 5.778967 95.76805 37.86543 137212.3 26 0.181085 5.724341 61.69456 30.6116 -9953.39 27 -0.51609 2.935652 109.5618 29.99144 -22535.3 28 -0.65776 2.368975 66.54915 22.74708 -169509 29 0.467112 6.868449 17.3875 22.95801 -165229 30 -0.10347 4.58614 28.54076 21.87101 -187283 31 -0.28159 3.873635 132.7862 37.21858 124088.9 32 0.215671 5.862686 100.8346 39.22092 164712.5 33 -0.39879 3.404844 153.0352 38.11374 142250 34 0.075346 5.301383 48.074 26.87515 -85758.6 35 0.10727 5.42908 77.24882 32.97342 37963.11
51
51
36 -1.07027 0.718925 86.16379 19.85382 -228207 37 0.549071 7.196284 139.9296 53.58913 456214.8 38 0.431603 6.726413 51.2168 30.68115 -8542.42 39 -0.09807 4.607739 147.7863 43.11443 243704 40 -0.69312 2.227535 104.0817 26.68361 -89644.6 41 0.720718 7.882874 19.64739 24.91783 -125469 42 -0.52738 2.890469 53.67027 22.42078 -176129 43 -0.06718 4.731272 85.0789 32.34901 25295.11 44 -0.2085 4.166005 46.29284 24.13256 -141400 45 0.354143 6.416573 92.14386 39.22628 164821.2 46 0.055809 5.223236 58.03419 28.64046 -49943.9 47 -0.46738 3.130461 111.8659 31.04799 -1100.05 48 -0.73816 2.047369 93.28242 24.8395 -127058 49 -0.53983 2.840685 96.63352 27.93762 -64203.2 50 -0.83657 1.653734 113.737 25.55865 -112468 51 0.011686 5.046744 49.76883 26.64755 -90376.2 52 -0.35621 3.575153 125.0916 34.74789 73963.49 53 -0.26709 3.931622 104.7063 32.98205 38138.32 54 -0.80571 1.777173 64.68256 20.99026 -205151 55 -0.25179 3.992828 76.23383 28.60526 -50658.1 56 0.114416 5.457665 31.35745 23.86276 -146874 57 -0.41822 3.327135 44.80873 22.15079 -181606 58 -0.16171 4.353144 51.92192 25.49021 -113856 59 -0.4478 3.208817 95.06784 28.98362 -42981.9 60 -0.11447 4.542103 70.31933 29.1567 -39470.6 61 0.590894 7.363576 108.4223 46.45435 311464.2 62 -0.04915 4.803396 36.6258 23.70666 -150041 63 -0.13121 4.47517 228.228 56.50328 515336.9 64 -0.84268 1.629292 119.9308 26.07675 -101957 65 -0.17774 4.289051 63.22651 27.26075 -77935.6 66 -0.2419 4.032391 97.4099 32.1637 21535.54 67 -0.19304 4.227823 102.9654 33.77226 54169.92 68 -0.31475 3.740993 121.9142 34.95115 78087.31 69 0.027628 5.110511 41.34747 25.17799 -120191 70 -0.00638 4.974475 61.02541 28.59985 -50767.8 71 0.074813 5.299251 168.1759 50.42934 392108.8 72 0.167952 5.671808 78.85104 34.01401 59074.64 73 0.876345 8.505379 130.0185 56.80787 521516.4 74 -0.30126 3.794974 164.6642 41.8564 218181.1 75 0.481129 6.924518 74.93782 36.79808 115557.9 76 -0.60293 2.588292 -14.4811 13.24711 -362244 77 -0.07685 4.692617 142.3177 42.53647 231978.3 78 0.323522 6.294088 80.21285 36.16492 102712.2 79 0.049092 5.196367 116.7473 39.93254 179149.7
52
52
80 0.391733 6.566934 124.2045 47.09045 324369.5 81 0.506403 7.025613 127.4007 49.77619 378857.8 82 0.647272 7.589089 177.7241 65.09517 689649.3 83 -0.34571 3.617172 67.47869 26.18819 -99695.6 84 0.311703 6.246812 87.93683 37.73167 134498.5 85 0.413713 6.654852 89.97998 39.51561 170691 86 -0.05331 4.786769 24.34059 21.4327 -196175 87 -0.02132 4.914725 134.8102 42.17295 224603.3 88 -0.38576 3.456948 98.95663 30.40015 -14243.3 89 -0.21773 4.129074 131.0421 38.05372 141032.3 90 -0.56524 2.739034 33.38566 19.46776 -236040 91 0.784015 8.13606 151.4384 61.02104 606993.3 92 0.682643 7.730571 74.06468 38.88077 157811.4 93 -1.00785 0.968586 66.22222 19.01787 -245167 94 0.221046 5.884183 -19.4038 13.85001 -350013 95 0.141384 5.565535 105.6908 39.1524 163322.3 96 -0.92827 1.286927 13.56775 15.12062 -324234 97 0.339725 6.358901 78.25181 35.92004 97744.22 98 0.159975 5.6399 87.26709 35.64601 92184.63 99 0.249983 5.999933 69.48713 32.97794 38054.75 100 0.379602 6.518407 73.04364 35.20149 83166.14
53
53
54
APPENDIX B SIMULATION OUTPUT FOR SCENARIO 2
Iteration "Z" "F" X(depth)Yld tons/ac Profit
1 -0.14087 4.43652 27.37056 21.42314 -196369 2 0.262596 6.050384 -7.03804 16.58278 -294570 3 0.293122 6.172489 79.73242 35.69409 93160.07 4 0.809249 8.236997 90.78386 44.90021 279933.8 5 0.621738 7.486952 127.0605 51.65264 416927.1 6 -0.70749 2.170035 33.61624 18.48899 -255897 7 0.267268 6.069074 62.47645 31.64807 11074.38 8 0.136634 5.546537 15.47828 20.79619 -209088 9 0.003309 5.013236 74.95033 31.32271 4473.455 10 0.551755 7.207022 82.00634 39.35124 167356.3 11 0.098492 5.393969 50.70392 27.59713 -71111.1 12 -0.44181 3.232743 98.79799 29.59259 -30627.1 13 -0.49417 3.023337 47.72059 21.91815 -186326 14 -0.37266 3.509378 105.7294 31.59139 9924.574 15 -0.59678 2.612878 111.88 29.06041 -41423.9 16 -0.77456 1.90174 -3.76276 14.04302 -346097 17 0.028506 5.114025 29.78745 22.97404 -164904 18 -0.88057 1.477735 45.82674 18.42208 -257254 19 -0.62596 2.496178 73.8853 23.97809 -144534 20 -0.0305 4.877987 136.5986 42.34032 227998.8 21 -0.17373 4.305074 102.0514 33.89573 56674.89 22 -0.32012 3.719529 148.6384 38.98756 159978.1 23 -1.12519 0.499255 72.72188 18.23016 -261148 24 0.456235 6.824941 109.3593 44.65483 274955.6 25 0.194742 5.778967 86.70932 35.97347 98828.18 26 0.181085 5.724341 52.63584 28.73274 -48071.7 27 -0.51609 2.935652 100.5031 28.78128 -47087 28 -0.65776 2.368975 57.49042 21.6728 -191304 29 0.467112 6.868449 8.328775 20.8048 -208914 30 -0.10347 4.58614 19.48203 20.26508 -219864 31 -0.28159 3.873635 123.7275 35.7835 94973.99 32 0.215671 5.862686 91.77583 37.30889 125921.1 33 -0.39879 3.404844 143.9765 36.79107 115415.7 34 0.075346 5.301383 39.01527 25.0977 -121819 35 0.10727 5.42908 68.19009 31.16536 1281.158
55
55
36 -1.07027 0.718925 77.10507 19.1752 -241975 37 0.549071 7.196284 130.8709 51.35732 410935.6 38 0.431603 6.726413 42.15807 28.562 -51535.7 39 -0.09807 4.607739 138.7276 41.50332 211017.8 40 -0.69312 2.227535 95.02302 25.64324 -110752 41 0.720718 7.882874 10.58866 22.52138 -174088 42 -0.52738 2.890469 44.61154 21.22145 -200461 43 -0.06718 4.731272 76.02017 30.70828 -7992.1 44 -0.2085 4.166005 37.23411 22.62737 -171937 45 0.354143 6.416573 83.08514 37.18143 123335.3 46 0.055809 5.223236 48.97546 26.88176 -85624.5 47 -0.46738 3.130461 102.8071 29.79111 -26599.5 48 -0.73816 2.047369 84.2237 23.84233 -147288 49 -0.53983 2.840685 87.57479 26.75023 -88293 50 -0.83657 1.653734 104.6783 24.65588 -130783 51 0.011686 5.046744 40.7101 24.93116 -125198 52 -0.35621 3.575153 116.0329 33.38438 46300.63 53 -0.26709 3.931622 95.6476 31.53307 8741.292 54 -0.80571 1.777173 55.62383 20.05789 -224067 55 -0.25179 3.992828 67.1751 27.1416 -80352.9 56 0.114416 5.457665 22.29872 22.04784 -183695 57 -0.41822 3.327135 35.75 20.84675 -208063 58 -0.16171 4.353144 42.86319 23.94015 -145304 59 -0.4478 3.208817 86.00911 27.70796 -68862.6 60 -0.11447 4.542103 61.2606 27.56132 -71837.5 61 0.590894 7.363576 99.36356 44.18241 265371.2 62 -0.04915 4.803396 27.56707 22.04863 -183679 63 -0.13121 4.47517 219.1692 54.92395 483295.5 64 -0.84268 1.629292 110.8721 25.17984 -120153 65 -0.17774 4.289051 54.16778 25.72605 -109071 66 -0.2419 4.032391 88.35117 30.69055 -8351.71 67 -0.19304 4.227823 93.90672 32.25224 23331.91 68 -0.31475 3.740993 112.8555 33.54787 49617.67 69 0.027628 5.110511 32.28874 23.44632 -155323 70 -0.00638 4.974475 51.96668 26.9008 -85238.2 71 0.074813 5.299251 159.1172 48.65241 356058.4 72 0.167952 5.671808 69.79231 32.14775 21211.86 73 0.876345 8.505379 120.9598 54.26214 469868.7 74 -0.30126 3.794974 155.6054 40.44018 189448.9 75 0.481129 6.924518 65.87909 34.63143 71600.87 76 -0.60293 2.588292 -23.5398 12.12024 -385106 77 -0.07685 4.692617 133.2589 40.90501 198879.2 78 0.323522 6.294088 71.15412 34.14943 61822.11 79 0.049092 5.196367 107.6885 38.18028 143599.9
56
56
80 0.391733 6.566934 115.1458 45.00954 282152 81 0.506403 7.025613 118.3419 47.5853 334409 82 0.647272 7.589089 168.6653 62.76917 642459.2 83 -0.34571 3.617172 58.41996 24.81461 -127563 84 0.311703 6.246812 78.8781 35.72752 93838.36 85 0.413713 6.654852 80.92125 37.41362 128045.9 86 -0.05331 4.786769 15.28186 19.77865 -229732 87 -0.02132 4.914725 125.7514 40.48823 190423.6 88 -0.38576 3.456948 89.89791 29.06499 -41331.1 89 -0.21773 4.129074 121.9834 36.55739 110674.7 90 -0.56524 2.739034 24.32693 18.30474 -259635 91 0.784015 8.13606 142.3797 58.56387 557142.3 92 0.682643 7.730571 65.00595 36.52083 109933 93 -1.00785 0.968586 57.16349 18.27939 -260149 94 0.221046 5.884183 -28.4625 11.93282 -388909 95 0.141384 5.565535 96.63204 37.31162 125976.5 96 -0.92827 1.286927 4.509018 14.30581 -340765 97 0.339725 6.358901 69.19309 33.88902 56538.82 98 0.159975 5.6399 78.20836 33.7874 54477.07 99 0.249983 5.999933 60.4284 31.03299 -1404.32 100 0.379602 6.518407 63.98491 33.13221 41184.76
57
APPENDIX C SIMULATION OUTPUT FOR SCENARIO 3
Iteration "Z" "F" X(depth)Yld tons/ac Profit
1 -0.14087 4.43652 5.949285 17.71041 -271693 2 0.262596 6.050384 -28.4593 11.95494 -388460 3 0.293122 6.172489 58.31114 30.99701 -2134.3 4 0.809249 8.236997 69.36259 39.03248 160889.4 5 0.621738 7.486952 105.6392 46.21021 306511.2 6 -0.70749 2.170035 12.19497 16.06143 -305147 7 0.267268 6.069074 41.05518 27.00963 -83030.3 8 0.136634 5.546537 -5.94299 16.45405 -297182 9 0.003309 5.013236 53.52906 27.28296 -77484.9 10 0.551755 7.207022 60.58506 34.06754 60160.66 11 0.098492 5.393969 29.28264 23.34149 -157449 12 -0.44181 3.232743 77.37672 26.56245 -92102.7 13 -0.49417 3.023337 26.29932 19.00675 -245393 14 -0.37266 3.509378 84.30816 28.40439 -54733.4 15 -0.59678 2.612878 90.45869 26.38175 -95768.6 16 -0.77456 1.90174 -25.184 11.7676 -392261 17 0.028506 5.114025 8.366183 18.87714 -248022 18 -0.88057 1.477735 24.40547 16.38708 -298540 19 -0.62596 2.496178 52.46403 21.36561 -197536 20 -0.0305 4.877987 115.1774 38.37726 147596.3 21 -0.17373 4.305074 80.6301 30.25753 -17136.8 22 -0.32012 3.719529 127.2171 35.68139 92902.54 23 -1.12519 0.499255 51.30061 16.74999 -291178 24 0.456235 6.824941 87.93798 39.58779 172155.4 25 0.194742 5.778967 65.28805 31.49953 8060.889 26 0.181085 5.724341 31.21456 24.28978 -138211 27 -0.51609 2.935652 79.08184 25.91959 -105145 28 -0.65776 2.368975 36.06915 19.13244 -242843 29 0.467112 6.868449 -13.0925 15.71309 -312214 30 -0.10347 4.58614 -1.93924 16.46751 -296909 31 -0.28159 3.873635 102.3062 32.38994 26125.6 32 0.215671 5.862686 70.35456 32.78748 34190.73 33 -0.39879 3.404844 122.5552 33.66334 51960.25 34 0.075346 5.301383 17.594 20.89457 -207093 35 0.10727 5.42908 46.76882 26.88981 -85461
58
58
36 -1.07027 0.718925 55.68379 17.57048 -274532 37 0.549071 7.196284 109.4496 46.07971 303863.5 38 0.431603 6.726413 20.7368 23.55083 -153202 39 -0.09807 4.607739 117.3063 37.69351 133724.3 40 -0.69312 2.227535 73.60175 23.18308 -160663 41 0.720718 7.882874 -10.8326 16.85445 -289059 42 -0.52738 2.890469 23.19027 18.38539 -257999 43 -0.06718 4.731272 54.5989 26.82841 -86706.7 44 -0.2085 4.166005 15.81284 19.06804 -244149 45 0.354143 6.416573 61.66386 32.34595 25232.97 46 0.055809 5.223236 27.55419 22.72294 -169999 47 -0.46738 3.130461 81.38587 26.81896 -86898.5 48 -0.73816 2.047369 62.80242 21.48434 -195127 49 -0.53983 2.840685 66.15352 23.94239 -145258 50 -0.83657 1.653734 83.25703 22.52108 -174094 51 0.011686 5.046744 19.28883 20.87242 -207542 52 -0.35621 3.575153 94.6116 30.16007 -19114 53 -0.26709 3.931622 74.22633 28.10664 -60774.2 54 -0.80571 1.777173 34.20256 17.8531 -268798 55 -0.25179 3.992828 45.75383 23.68046 -150572 56 0.114416 5.457665 0.877447 17.75609 -270766 57 -0.41822 3.327135 14.32873 17.76309 -270624 58 -0.16171 4.353144 21.44192 20.2747 -219669 59 -0.4478 3.208817 64.58784 24.69138 -130063 60 -0.11447 4.542103 39.83933 23.78873 -148376 61 0.590894 7.363576 77.94229 38.80995 156374.6 62 -0.04915 4.803396 6.145801 18.12787 -263223 63 -0.13121 4.47517 197.748 51.18931 407527.1 64 -0.84268 1.629292 89.45081 23.0589 -163183 65 -0.17774 4.289051 32.74651 22.09695 -182699 66 -0.2419 4.032391 66.9299 27.20698 -79026.4 67 -0.19304 4.227823 72.48544 28.65785 -49591 68 -0.31475 3.740993 91.43423 30.22953 -17704.8 69 0.027628 5.110511 10.86747 19.35141 -238400 70 -0.00638 4.974475 30.54541 22.88303 -166751 71 0.074813 5.299251 137.6959 44.45048 270809.8 72 0.167952 5.671808 48.37104 27.73457 -68322.7 73 0.876345 8.505379 99.5385 48.24223 347736.8 74 -0.30126 3.794974 134.1842 37.09124 121505.4 75 0.481129 6.924518 44.45782 29.50793 -32344.8 76 -0.60293 2.588292 -44.9611 9.455519 -439168 77 -0.07685 4.692617 111.8377 37.04706 120609.2 78 0.323522 6.294088 49.73285 29.3834 -34871.1 79 0.049092 5.196367 86.26727 34.03669 59534.82
59
59
80 0.391733 6.566934 93.72455 40.0888 182320 81 0.506403 7.025613 96.92066 42.40447 229300.3 82 0.647272 7.589089 147.2441 57.26882 530868.3 83 -0.34571 3.617172 36.99869 21.56648 -193461 84 0.311703 6.246812 57.45683 30.9883 -2311.01 85 0.413713 6.654852 59.49998 32.44302 27202.42 86 -0.05331 4.786769 -6.13941 15.86732 -309085 87 -0.02132 4.914725 104.3302 36.50434 109598.5 88 -0.38576 3.456948 68.47663 25.90771 -105386 89 -0.21773 4.129074 100.5621 33.01899 38887.7 90 -0.56524 2.739034 2.905657 15.55455 -315431 91 0.784015 8.13606 120.9584 52.75338 439259 92 0.682643 7.730571 43.58468 30.94027 -3285.47 93 -1.00785 0.968586 35.74222 16.5331 -295578 94 0.221046 5.884183 -49.8838 7.399221 -480886 95 0.141384 5.565535 75.21077 32.9587 37664.54 96 -0.92827 1.286927 -16.9123 12.37901 -379856 97 0.339725 6.358901 47.77181 29.08624 -40900 98 0.159975 5.6399 56.78709 29.39231 -34690.4 99 0.249983 5.999933 39.00713 26.43375 -94713.6 100 0.379602 6.518407 42.56364 28.23899 -58089
60
APPENDIX D SIMULATION OUTPUT FOR SCENARIO 4
Iteration "Z" "F" X(depth)Yld tons/ac Profit
1 -0.14087 4.43652 3.475841 17.28171 -280390 2 0.262596 6.050384 -17.6044 14.30002 -340883 3 0.293122 6.172489 46.1128 28.32226 -56399.7 4 0.809249 8.236997 54.77228 35.0359 79806.66 5 0.621738 7.486952 78.20701 39.24061 165111.9 6 -0.70749 2.170035 8.295705 15.61955 -314112 7 0.267268 6.069074 31.95968 25.04014 -122987 8 0.136634 5.546537 -5.05496 16.63406 -293530 9 0.003309 5.013236 42.24638 25.15521 -120653 10 0.551755 7.207022 47.92933 30.94592 -3170.76 11 0.098492 5.393969 22.16669 21.92781 -186130 12 -0.44181 3.232743 60.72192 24.20654 -139899 13 -0.49417 3.023337 19.69937 18.10973 -263591 14 -0.37266 3.509378 65.57358 25.6171 -111282 15 -0.59678 2.612878 69.60374 23.77391 -148677 16 -0.77456 1.90174 -16.1404 12.72824 -372771 17 0.028506 5.114025 5.318517 18.29427 -259848 18 -0.88057 1.477735 18.13991 15.79186 -310616 19 -0.62596 2.496178 41.37789 20.01357 -224966 20 -0.0305 4.877987 82.48883 32.3297 24903.42 21 -0.17373 4.305074 63.03677 27.26948 -77758.4 22 -0.32012 3.719529 86.58802 29.41068 -34317.7 23 -1.12519 0.499255 40.4265 15.99861 -306422 24 0.456235 6.824941 67.98594 34.86828 76406.07 25 0.194742 5.778967 51.63238 28.64748 -49801.6 26 0.181085 5.724341 23.76982 22.74567 -169537 27 -0.51609 2.935652 61.94307 23.63001 -151596 28 -0.65776 2.368975 27.80888 18.15285 -262717 29 0.467112 6.868449 -9.63148 16.53575 -295524 30 -0.10347 4.58614 -2.29595 16.40427 -298192 31 -0.28159 3.873635 76.4983 28.30146 -56821.5 32 0.215671 5.862686 55.52589 29.65757 -29308.7 33 -0.39879 3.404844 85.17059 28.20479 -58782.8 34 0.075346 5.301383 12.59395 19.9135 -226997 35 0.10727 5.42908 36.69784 24.87971 -126242
61
61
36 -1.07027 0.718925 43.99566 16.69489 -292296 37 0.549071 7.196284 80.02718 38.83084 156798.5 38 0.431603 6.726413 15.13879 22.24126 -179771 39 -0.09807 4.607739 83.31895 31.64879 11089.11 40 -0.69312 2.227535 57.96023 21.38671 -197108 41 0.720718 7.882874 -8.23757 17.54096 -275131 42 -0.52738 2.890469 17.14267 17.58472 -274243 43 -0.06718 4.731272 43.11628 24.74866 -128901 44 -0.2085 4.166005 11.16465 18.2957 -259818 45 0.354143 6.416573 48.78549 29.43888 -33745.7 46 0.055809 5.223236 20.73577 21.39918 -196855 47 -0.46738 3.130461 63.56523 24.34639 -137062 48 -0.73816 2.047369 49.68489 20.04039 -224422 49 -0.53983 2.840685 52.30501 22.12717 -182086 50 -0.83657 1.653734 64.85773 20.68745 -211295 51 0.011686 5.046744 13.96295 19.86331 -228015 52 -0.35621 3.575153 72.15883 26.78051 -87678.6 53 -0.26709 3.931622 58.42244 25.57873 -112060 54 -0.80571 1.777173 26.25471 17.03507 -285394 55 -0.25179 3.992828 35.85858 22.08163 -183009 56 0.114416 5.457665 -0.28206 17.52378 -275479 57 -0.41822 3.327135 9.981866 17.13734 -283319 58 -0.16171 4.353144 15.7133 19.29446 -239556 59 -0.4478 3.208817 51.08607 22.79004 -168637 60 -0.11447 4.542103 30.94818 22.22287 -180144 61 0.590894 7.363576 61.12885 34.59313 70823.78 62 -0.04915 4.803396 3.624508 17.66639 -272586 63 -0.13121 4.47517 92.73595 32.88123 36092.71 64 -0.84268 1.629292 68.96274 21.03036 -204338 65 -0.17774 4.289051 25.04319 20.79188 -209176 66 -0.2419 4.032391 52.90587 24.92636 -125296 67 -0.19304 4.227823 57.12919 26.08115 -101867 68 -0.31475 3.740993 70.21656 26.94273 -84387.4 69 0.027628 5.110511 7.256011 18.66104 -252406 70 -0.00638 4.974475 23.21417 21.50799 -194648 71 0.074813 5.299251 89.05367 34.90898 77231.89 72 0.167952 5.671808 38.01983 25.60203 -111588 73 0.876345 8.505379 74.99924 41.3461 207828 74 -0.30126 3.794974 88.33166 29.92277 -23928.5 75 0.481129 6.924518 34.7852 27.19444 -79280.7 76 -0.60293 2.588292 -22.9999 12.1874 -383744 77 -0.07685 4.692617 81.09388 31.51015 8276.34 78 0.323522 6.294088 39.1404 27.02668 -82684.3 79 0.049092 5.196367 66.88735 30.28797 -16519.2
62
62
80 0.391733 6.566934 71.62503 35.01225 79327.03 81 0.506403 7.025613 73.51675 36.74413 114463.3 82 0.647272 7.589089 90.55837 42.71363 235572.6 83 -0.34571 3.617172 28.58303 20.2904 -219350 84 0.311703 6.246812 45.42643 28.3267 -56309.5 85 0.413713 6.654852 47.06445 29.55748 -31339.5 86 -0.05331 4.786769 -5.18694 16.04124 -305557 87 -0.02132 4.914725 77.5487 31.52358 8548.78 88 -0.38576 3.456948 54.0955 23.78808 -148389 89 -0.21773 4.129074 75.56196 28.88944 -44892.7 90 -0.56524 2.739034 1.201646 15.33578 -319869 91 0.784015 8.13606 84.63641 42.90109 239375.8 92 0.682643 7.730571 34.06107 28.45922 -53620.9 93 -1.00785 0.968586 27.53662 15.86416 -309149 94 0.221046 5.884183 -24.0274 12.87146 -369865 95 0.141384 5.565535 59.14679 29.69442 -28561.3 96 -0.92827 1.286927 -11.8668 12.83283 -370649 97 0.339725 6.358901 37.52584 26.78903 -87505.8 98 0.159975 5.6399 44.88694 26.95071 -84225.5 99 0.249983 5.999933 30.25548 24.55474 -132835 100 0.379602 6.518407 33.21341 26.10313 -101421
63
LIST OF REFERENCES
Aillery, M., R. Shoemaker, and M. Caswell. “Agriculture and Ecosystem Restoration in South Florida: Assessing Trade-offs from Water-Retention Development in the Everglades Agricultural Area”. Amer. J. Agr. Econ 83 (1) (February 2001):183-195.
Alchian, A.A. and H. Demsetz. “Production, Information Costs, and Economic Organization.” Amer. Econ. Rev. 62(December 1972):777-795.
Alvarez, J. and T. J. Schuneman. “Costs and Returns for Sugarcane Production on Muck Soils in Florida, 1990-91.” Economics Information Report EI 91-3. Institute of Food and Agricultural Sciences, The University of Florida, (revised) June 1998.
Apland, J., Grainger, C., and Strock, J. “Modeling Agricultural Production Considering Water Quality and Risk,” Department of Agricultural and Applied Economics, University of Minnesota, Staff Paper P04-13. November, 2004.
Boggess, W., R. Lacewell and D. Zilberman. “Economics of Water Use in Agriculture," in Agricultural and Environmental Resource Economics, eds. Gerald A. Carlson, David Zilberman and John A. Miranowski, Oxford Series in Biological Resource Management, Oxford University Press, New York, 1993:
Bottcher, A.B. and F.T. Izuno, eds. Everglades Agricultural Area(EAA): Water, Soil, Crop, and Environmental Management: Gainesville, FL: University Press of Florida, 1994.
Brue, S.L. “Retrospectives: The Law of Diminishing Returns. J. of Econ. Persp. 7(Summer, 1993):185-192.
Coase, R. H. “The Nature of the Firm.” Economica. 4(November 1937):386-405.
Diewert, W.E. “Functional Forms for Profit and Transformation Functions,” J. Econ Theory 6(1973):284-316.
Diewert, W.E. and T.J. Wales. “Flexible Functional Forms and Global Curvature Conditions,” Econometrica 55(1987):43-68.
Douglas, M.S. Everglades: River of Grass: St. Simons Island, GA: Mockingbird Press, 1947 (p.18).
64
64
Edgeworth, F.Y. “Contributions to the Theory of Railway Rates.” The Econ. J. 21(September 1911):346-370.
Elbadawi, I., R. Gallant, and G. Souza. “An Elasticity Can Be Estimated Consostently Without A Priori Knowledge of Functional Form,” Econometrica 51(1983):1731-51.
Everglades Research and Education Center (EREC), 2005. “EREC Weather Station. http://erec.ifas.ufl.edu/WD/EWDMAIN.HTM. Accessed July 8, 2004.
Gallant, R. “On the Bias in Flexible Functional Forms and an Essentially Unbiased Form,” J. Econometrics 15(1981):211-45.
Gallant, R. “Unbiased Determination of Production Technologies,” J. Econometrics 20(1982):285-323.
Glaz, B. “Sugarcane Emergence after Long Duration under Water” Soil Crop Sci. Florida Proc. 62(2003):51-57.
Glaz, B. and R. Cherry. “Wireworm Effects on Sugarcane Emergence after Short-Duration Flood Applied at Planting” J. Entomol. Sci. 38 (July 2003):449-456.
Glaz, B., S. Edme, J. Miller, S. Milligan, and D. Holder. “Sugarcane Cultivar Response to High Water Tables in the Everglades” Agron J. 94(2002):624-629.
Just, R.E. and R.D. Pope. “Production Function Estimation and Related Risk Considerations.” Amer. J. Agr. Econ 61(May 1979):276-284.
Kim, C.S. and Glenn D. Schaible. “Economic Benefits Resulting From Irrigation Water Use: Theory and an Application to Groundwater Use” Environ. and Res. Econ. 17(September 2000):73-87.
Lang, T.A., S. H. Daroub and R. S. Lentini. “Water Management for Florida Sugarcane Production” Circular SS-AGR-231. Institute of Food and Agricultural Sciences, The University of Florida, May 2002.
Omary, M. and F Izuno. “Evaluation of Sugarcane Evapotranspiration from Water Table Data in the Everglades Agricultural Area.” Agric Water Manage. 27(1995):309-319.
Paris, Q. and M.R. Caputo. “A Primal-Dual Estimator of Production and Cost Functions Within an Errors-in-Variables Context” Department of Agricultural and Resource Economics University of California, Davis Working Paper No. 04-008. September, 2004.
Redman, J.C. and S.Q.Allen. “Some Interrelationships of Economic and Agronomic Concepts.” J. of Farm Econ. 36 (August 1954):453-465.
65
65
South Florida Water Management District (SFWMD). “Florida Forever Work Plan, 2004 Annual Update.” West Palm Beach, FL, February 2004.
South Florida Water Management District (SFWMD). “Everglades Agricultural Area Storage Reservoir Phase 1 Existing Flood Control Conditions Documentation.” West Palm Beach, FL, 2004.
Thompson, Gary D. “Choice of Flexible Functional Forms: Review and Appraisal” Western J. of Agr. Econ., 13(December 1988):169-183.
U.S. Army Corps of Engineers (USACE) and South Florida Water Management District (SFWMD). “Central and Southern Florida Project- Comprehensive Review Study: Integrated Feasibility Report and Programmatic Environmental Impact Statement .” USACE Jacksonville, District, FL, April 1999.
U.S. Army Corps of Engineers (USACE) and South Florida Water Management District (SFWMD). “Environmental and Economic Equity Program Management Plan.” USACE Jacksonville District, FL, August 2001.
U.S. Army Corps of Engineers (USACE) and South Florida Water Management District (SFWMD). “Regional Economic Impact- Everglades Agricultural Area Storage Reservoirs- Phase 1.” USACE Jacksonville District, FL, October 2003.
U.S. Geological Survey, 2002. “Land and People: Finding a Balance; Everglades” http://interactive2.usgs.gov/learningweb/pdf/landpeople/evergladesst.pdf. Accessed February 6, 2005.
Weisskoff, R.. The Economics of Everglades Restoration: Missing Pieces in the Future of South Florida. Northhampton, MA: Edward Elgar Publishing, 2004.
66
BIOGRAPHICAL SKETCH
Jennie Varela was born and raised in St. Petersburg, Florida. She was a graduate of
the International Baccalaureate Program at St. Petersburg High School where she was
named a National Merit Scholar. She continued her education at the University of
Florida in Gainesville, Florida. She graduated with a Bachelor of Science in food and
resource economics in May 2002 and completed a Master of Science program in the same
department in 2005. She has accepted a position as an Agricultural Economist in Dairy
Programs for the Agricultural Marketing Service of USDA.
